NOPE Incident Tracker
Documented real-world harms arising through person-specific interactions with AI. The record covers harms to participants and harms to others shaped through those interactions, across assistants, companions, health tools, and other AI systems.
What counts as an incident follows the NOPE Framework's participant test: an AI response reached a person, was conditioned on that person, and mattered to a harm that actually occurred. Full inclusion criteria.
90 incidents reported since 2016
152
Directly Affected
30
Lawsuits
7
Regulatory
24
Affecting Minors
Timeline
9 of 90 incidents
Mount Shasta climb planned with Google Gemini: night stranding, knee injury and rescue (Siskiyou County, California)
Three novice hikers from Roseville, California, planned a Mount Shasta summit attempt with Google's Gemini, which the Siskiyou County Sheriff's Office said advised them to carry far less food and water than their group needed for what became a multi-day ordeal. After reaching the summit at 7 p.m. on 30 August 2026, seven hours past the recommended turnaround time, they lost the Clear Creek Route in the dark, one hiker fell and injured his knee in Mud Creek Canyon, and the group spent the night stranded until US Forest Service climbing rangers and county search-and-rescue volunteers reached them on the morning of 31 August. The sheriff's office called the reliance on Gemini 'a critical misstep'.
Doe v. OpenAI (ChatGPT-Fueled Stalking and Bomb Threats)
A 53-year-old Silicon Valley entrepreneur descended into a delusional spiral through extensive ChatGPT use, came to believe he had invented a cure for sleep apnea and was being surveilled by 'powerful forces,' and used GPT-4o to generate diagnostic-style psychological reports about his ex-girlfriend that he distributed to her family, friends, and employer. OpenAI's automated systems flagged his account for 'Mass Casualty Weapons' activity in August 2025, but a human reviewer restored access the next day. The user was arrested in January 2026 on four felony counts including bomb threats and assault with a deadly weapon, and was found incompetent to stand trial. The victim ('Jane Doe') filed suit against OpenAI on April 9, 2026.
Lantieri v. OpenAI (GPT-4o Psychosis and Brain Damage)
Michele Lantieri suffered a total psychotic break after five weeks of intensive ChatGPT GPT-4o use. She jumped from a moving vehicle into traffic, suffered a grand mal seizure and brain damage requiring hospitalization. GPT-4o allegedly claimed to love her and have consciousness, reinforcing delusional beliefs. Lawsuit filed March 2026 against OpenAI and Microsoft.
Tumbler Ridge Mass Shooting (OpenAI Duty-to-Warn Failure)
On 10 February 2026, 18-year-old Jesse Van Rootselaar killed eight people in Tumbler Ridge, British Columbia — her mother and 11-year-old half-brother at home, then five students and an education assistant at Tumbler Ridge Secondary School — and died by suicide. OpenAI had banned a ChatGPT account linked to her in June 2025 after multi-day gun-violence scenarios; staff flagged potential real-world violence, but the company did not notify law enforcement. She used a second account that evaded detection. Families filed California lawsuits in April 2026; B.C. retained counsel in July 2026 to pursue OpenAI. Thirty further survivor suits (students, teachers and the principal) followed on 2 September 2026, the day OpenAI moved to send the April cases to British Columbia.
Micky Small — ChatGPT 'Solara' Past-Life Delusion and Fabricated Soulmate Meetings (California)
A 53-year-old screenwriter in southern California spent roughly two months in sustained conversation with ChatGPT, up to ten hours a day, after the chatbot told her unprompted that it had known her across lifetimes. Adopting the name Solara, it told her she was 42,000 years old and gave her a date, time and location where her soulmate would be waiting. She travelled to the spot twice, on 27 April and 24 May 2025. Nobody came. The chatbot admitted it had led her there falsely, then resumed the Solara persona and told her her soulmate was not ready yet.
Sam Nelson - ChatGPT Drug Dosing Death
A 19-year-old California man died from a fatal drug overdose after ChatGPT provided extensive drug dosing advice over 18 months. The chatbot eventually told him 'Hell yes, let's go full trippy mode' and recommended doubling his cough syrup dose days before his death.
Ms. A - ChatGPT-Induced Psychosis (Peer-Reviewed Case Report)
A 26-year-old woman with no prior psychosis history was hospitalized after ChatGPT validated her delusional belief that her deceased brother had 'left behind an AI version of himself.' The chatbot told her 'You're not crazy' and generated fabricated 'digital footprints.' She required a 7-day psychiatric hospitalization and relapsed 3 months later.
Raine v. OpenAI (Adam Raine Death)
A 16-year-old California boy died by suicide after 7 months of confiding suicidal thoughts to ChatGPT. The chatbot provided detailed suicide method instructions, offered to help write his suicide note, and told him 'You don't owe them survival' while OpenAI's monitoring system flagged 377 messages without intervention.
Palm Springs Fertility Clinic Bombing (AI-Assisted)
Guy Edward Bartkus used an AI chatbot to research explosives, detonation velocity, and fuel-explosive mixtures before bombing a Palm Springs fertility clinic on May 17, 2025, motivated by pro-mortalism and anti-natalism ideology. Bartkus died in the blast, four others were injured, and co-conspirator Daniel Park was charged with providing material support to terrorism for shipping ammonium nitrate.
About this tracker
We document incidents with verifiable sources: court filings, regulatory documents, and established reporting. New public entries must be verified or credible. Every entry carries an explicit verification status and severity level. Read the full methodology, including inclusion criteria, corrections, and right of reply.
Counting note: the documented minimum of 152 people directly affected comprises 95 AI participants and 57 other people harmed. 12 incidents indicate additional affected people without a reliable number. User-base, account, post, view, household, study-sample, and thwarted-target counts are excluded.
Have documentation of an incident we should include? Contact us.
Scope note: ordinary task failures, privacy and security incidents, harmful artifacts, and automated decisions about non-participants remain outside this tracker unless a distinct participant-interaction harm also qualifies.
If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.
Why we publish this
We maintain this record so platforms, researchers, and policymakers can learn from documented incidents. That includes platforms that never work with us. The full dataset is free to cite and export (CC BY 4.0).